Weeks of work on a client's website redesign.
I'd analyzed their business. Studied their competitors. Interviewed their team. Built something I was certain would solve their problem.
Thirty minutes into the presentation, they stopped me.
"We don't need this."
That was the whole conversation.
I'd spent weeks building something nobody wanted. Not because I was lazy or incompetent. Because I was confident.
My assumptions felt like facts. My bias became my product.
Classic Founder Effect.
In population genetics, the founder effect describes what happens when a small group establishes a new population.
Their genetic quirks — including blind spots — become the foundation for everyone who comes after.
Same with products. The founder's assumptions become features. Their blind spots become bugs that nobody notices until it's too late.
After that meeting, I started noticing this pattern everywhere.
Founders building features nobody asked for. Startups pivoting after the money ran out. Products dying not from bad execution, but from bad assumptions.
The problem isn't that founders don't do research. It's that research methods lie to them.
Surveys capture what people say, not what they do. Focus groups capture performance, not truth. Personas are static — they don't change their minds or influence each other.
What if you could test your idea against a population that actually behaves like humans?
Not perfect predictions. But a map of where the friction is. Before you waste months building the wrong thing.
Synthetic populations for idea validation.
AI agents that remember past conversations. That form opinions and change them. That influence each other through social networks. That age and pass values to the next generation.
Throw your idea into this population. Watch what happens.
Not what you hope happens. What actually happens.
We're building this.
Get early access when we launch.